From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jack Stone Subject: Re: [PATCH] fs/eventpoll.c: fix compilation warning Date: Fri, 14 Jan 2011 10:55:14 +0000 Message-ID: <4D302B92.902@fastmail.fm> References: <1294989352-2003-1-git-send-email-viresh.kumar@st.com> <2426ea8f-7e52-4329-bf20-884c2b996f98@email.android.com> <4D30198D.9080206@st.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Cc: "linux-kernel@vger.kernel.org" , "linux-fsdevel@vger.kernel.org" To: viresh kumar Return-path: Received: from out1.smtp.messagingengine.com ([66.111.4.25]:46391 "EHLO out1.smtp.messagingengine.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751152Ab1ANKzR (ORCPT ); Fri, 14 Jan 2011 05:55:17 -0500 In-Reply-To: <4D30198D.9080206@st.com> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On 14/01/2011 09:38, viresh kumar wrote: > On 01/14/2011 03:03 PM, Jack Stone wrote: >>>> - long slack; >>>> + long slack = 0; >>>> wait_queue_t wait; >>>> struct timespec end_time; >>>> ktime_t expires, *to = NULL; >> I don't think this is the correct fix. This function is fine unless timeout is negative. >> >> If a negative timeout is possible then this function will create timer far in the future. >> I'll leave it up to the maintainer how to solve that one. The two solutions I can see are making >> timeout unsigned or extending the bottom case of the if to <=0. >> >> Either way we should use uninitalized_var() rather than setting it to zero. > > Using uninitialized var gives compilation warning, How should we fix that. > I meant something like - long slack; + long uninitialized_var(slack); I just had a go at generating this warning and couldn't. I've tried GCC 4.4.5 (From Fedora 13) and latest sparse. Thanks, Jack